| *** NM has quit IRC | 00:05 | |
| *** manous has quit IRC | 00:32 | |
| *** diogogmt has quit IRC | 00:46 | |
| *** manous has joined #openstack-swift | 00:48 | |
| *** m_kazuhiro has joined #openstack-swift | 00:53 | |
| *** nadeem has joined #openstack-swift | 00:56 | |
| *** klrmn has joined #openstack-swift | 01:03 | |
| *** nadeem has quit IRC | 01:08 | |
| *** nadeem has joined #openstack-swift | 01:09 | |
| *** jamielennox is now known as jamielennox|away | 01:09 | |
| *** nadeem has quit IRC | 01:13 | |
| openstackgerrit | Matthew Oliver proposed openstack/swift: Add install-guide for swift https://review.openstack.org/330070 | 01:21 |
|---|---|---|
| mattoliverau | ^ this is just correcting a merge conflict now that it depends on the move of the api-ref. | 01:24 |
| *** manous has quit IRC | 01:31 | |
| *** manous has joined #openstack-swift | 01:43 | |
| *** jamielennox|away is now known as jamielennox | 01:43 | |
| *** ManojK has joined #openstack-swift | 01:53 | |
| *** jamielennox is now known as jamielennox|away | 01:53 | |
| *** ManojK has quit IRC | 01:59 | |
| zaitcev | Hugh came back to me about the xattr in /tmp. Unfortunately, my patch needs a significant development... http://marc.info/?l=linux-mm&m=146767747708574&w=2 | 02:03 |
| *** diogogmt has joined #openstack-swift | 02:07 | |
| *** asettle has joined #openstack-swift | 02:13 | |
| *** asettle has quit IRC | 02:19 | |
| *** furlongm has joined #openstack-swift | 02:20 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Fix EC ring validation at ring reload https://review.openstack.org/268993 | 02:36 |
| *** jamielennox|away is now known as jamielennox | 02:43 | |
| *** smile-luobin_ has joined #openstack-swift | 03:06 | |
| *** rcernin has joined #openstack-swift | 03:08 | |
| *** takashi has joined #openstack-swift | 03:13 | |
| *** manous has quit IRC | 03:20 | |
| *** janonymous has joined #openstack-swift | 03:23 | |
| *** ManojK has joined #openstack-swift | 03:53 | |
| *** vint_bra has joined #openstack-swift | 04:00 | |
| *** klrmn has quit IRC | 04:07 | |
| *** vint_bra has quit IRC | 04:08 | |
| *** vint_bra has joined #openstack-swift | 04:08 | |
| *** psachin has joined #openstack-swift | 04:29 | |
| *** ManojK has quit IRC | 04:29 | |
| *** kei_yama has quit IRC | 04:44 | |
| *** kei_yama has joined #openstack-swift | 04:46 | |
| *** vint_bra has quit IRC | 04:59 | |
| *** kei_yama_ has joined #openstack-swift | 05:02 | |
| *** kei_yama has quit IRC | 05:05 | |
| *** sheel has joined #openstack-swift | 05:07 | |
| *** hseipp has joined #openstack-swift | 05:07 | |
| *** kei_yama has joined #openstack-swift | 05:08 | |
| *** kei_yama_ has quit IRC | 05:09 | |
| *** ppai has joined #openstack-swift | 05:14 | |
| *** kei_yama has quit IRC | 05:20 | |
| *** kei_yama has joined #openstack-swift | 05:21 | |
| openstackgerrit | yuyafei proposed openstack/python-swiftclient: Add __ne__ built-in function https://review.openstack.org/337427 | 05:30 |
| openstackgerrit | yuyafei proposed openstack/swift: Add __ne__ built-in function https://review.openstack.org/337428 | 05:37 |
| *** SkyRocknRoll has joined #openstack-swift | 05:39 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: WIP: Force close backend socket at putter close https://review.openstack.org/337431 | 05:44 |
| kota_ | acoes:^^ patch 337431 might be better way to close backend connection on Putters. | 05:44 |
| patchbot | kota_: https://review.openstack.org/#/c/337431/ - swift - WIP: Force close backend socket at putter close | 05:44 |
| *** hseipp has quit IRC | 05:46 | |
| *** hseipp has joined #openstack-swift | 05:46 | |
| *** hseipp has left #openstack-swift | 05:47 | |
| *** asettle has joined #openstack-swift | 05:48 | |
| *** ChubYann has quit IRC | 05:50 | |
| openstackgerrit | Cheng Li proposed openstack/python-swiftclient: Add python version constraint python>=2.7 https://review.openstack.org/334333 | 05:53 |
| *** asettle has quit IRC | 05:54 | |
| *** rcernin has quit IRC | 06:01 | |
| *** ppai has quit IRC | 06:09 | |
| *** smile-luobin_ has quit IRC | 06:12 | |
| *** ppai has joined #openstack-swift | 06:24 | |
| *** hogepodge has quit IRC | 06:25 | |
| *** hogepodge has joined #openstack-swift | 06:27 | |
| *** rcernin has joined #openstack-swift | 06:32 | |
| *** pcaruana has joined #openstack-swift | 06:46 | |
| *** tesseract- has joined #openstack-swift | 07:09 | |
| openstackgerrit | yuyafei proposed openstack/swift: Remove white space between print and () https://review.openstack.org/337485 | 07:14 |
| *** rledisez has joined #openstack-swift | 07:17 | |
| *** hseipp has joined #openstack-swift | 07:26 | |
| *** ppai has quit IRC | 07:37 | |
| *** chlong has quit IRC | 07:52 | |
| *** ppai has joined #openstack-swift | 07:52 | |
| *** chlong has joined #openstack-swift | 08:04 | |
| *** ppai has quit IRC | 08:09 | |
| openstackgerrit | Merged openstack/swift: Adds migrated API reference files https://review.openstack.org/312315 | 08:17 |
| *** takashi has quit IRC | 08:19 | |
| *** asettle has joined #openstack-swift | 08:19 | |
| *** ppai has joined #openstack-swift | 08:22 | |
| *** wanghua has quit IRC | 08:25 | |
| *** ppai has quit IRC | 08:29 | |
| *** acoles_ is now known as acoles | 08:38 | |
| acoles | kota_: re patch 215276 yes please push rebase (sorry, been offline) | 08:39 |
| patchbot | acoles: https://review.openstack.org/#/c/215276/ - swift - Enable object server to return non-durable data | 08:39 |
| *** ppai has joined #openstack-swift | 08:42 | |
| acoles | kota_: yeah, I was taking a rest :) | 08:45 |
| *** takashi has joined #openstack-swift | 08:50 | |
| asettle | mattoliverau: you around? :) | 08:54 |
| openstackgerrit | Christian Schwede proposed openstack/swift: Add support to increase object ring partition power https://review.openstack.org/337297 | 08:59 |
| *** ppai has quit IRC | 09:04 | |
| *** ouchkernel has quit IRC | 09:04 | |
| *** sanchitmalhotra has joined #openstack-swift | 09:06 | |
| *** ouchkernel has joined #openstack-swift | 09:09 | |
| *** sanchitmalhotra has quit IRC | 09:13 | |
| *** ppai has joined #openstack-swift | 09:16 | |
| *** asettle has quit IRC | 09:24 | |
| *** asettle has joined #openstack-swift | 09:24 | |
| kota_ | acoles: o/ | 09:27 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/swift: Enable object server to return non-durable data https://review.openstack.org/215276 | 09:28 |
| kota_ | done | 09:28 |
| *** asettle has quit IRC | 09:30 | |
| *** ppai has quit IRC | 09:31 | |
| *** asettle has joined #openstack-swift | 09:41 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Only use Timeout if we don't have a final_resp https://review.openstack.org/336671 | 10:05 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/swift: [WIP] Testing latest u-c https://review.openstack.org/318441 | 10:10 |
| *** ppai has joined #openstack-swift | 10:10 | |
| *** m_kazuhiro has quit IRC | 11:10 | |
| *** cdelatte has joined #openstack-swift | 11:17 | |
| *** asettle has quit IRC | 11:20 | |
| *** asettle has joined #openstack-swift | 11:21 | |
| *** ppai has quit IRC | 11:24 | |
| *** asettle has quit IRC | 11:26 | |
| *** ppai has joined #openstack-swift | 11:37 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Add encrypter and decrypter links to middleware.rst https://review.openstack.org/336618 | 11:39 |
| *** asettle has joined #openstack-swift | 11:47 | |
| *** SkyRocknRoll has quit IRC | 12:04 | |
| *** raildo-afk is now known as raildo | 12:07 | |
| *** chlong has quit IRC | 12:12 | |
| *** openstackgerrit has quit IRC | 12:19 | |
| *** openstackgerrit has joined #openstack-swift | 12:19 | |
| openstackgerrit | Merged openstack/swift: Remove white space between print and () https://review.openstack.org/337485 | 12:37 |
| *** chlong has joined #openstack-swift | 12:43 | |
| tdasilva | good morning | 12:45 |
| *** kei_yama has quit IRC | 12:55 | |
| openstackgerrit | Merged openstack/swift: Only use Timeout if we don't have a final_resp https://review.openstack.org/336671 | 12:57 |
| *** takashi has quit IRC | 13:02 | |
| rcernin | zaitcev: what would be min requirements for standalone swift storage nodes? | 13:03 |
| *** klamath has joined #openstack-swift | 13:08 | |
| *** klamath has quit IRC | 13:09 | |
| *** klamath has joined #openstack-swift | 13:09 | |
| *** NM has joined #openstack-swift | 13:11 | |
| *** _JZ_ has joined #openstack-swift | 13:12 | |
| *** wer has joined #openstack-swift | 13:20 | |
| *** sheel has quit IRC | 13:25 | |
| *** ManojK has joined #openstack-swift | 13:30 | |
| *** raildo has quit IRC | 13:34 | |
| *** ppai has quit IRC | 13:35 | |
| *** links has joined #openstack-swift | 13:38 | |
| *** links has quit IRC | 13:45 | |
| *** diogogmt has quit IRC | 13:47 | |
| *** diogogmt has joined #openstack-swift | 13:49 | |
| psachin | ping #openstack-swift, can we use nfs mount point as swift node ? I can't find the configuration | 13:51 |
| *** ManojK has quit IRC | 13:53 | |
| *** links has joined #openstack-swift | 13:55 | |
| *** links has quit IRC | 13:59 | |
| *** diogogmt has quit IRC | 14:02 | |
| rcernin | psachin: something like this ? https://github.com/openstack/swiftonfile | 14:10 |
| rcernin | psachin: or what exactly are you trying to do, what exactly do you mean by use nfs mount as swift node? | 14:11 |
| *** permalac has quit IRC | 14:14 | |
| openstackgerrit | Merged openstack/swift: Stop digging for publicly_accessible ourselves https://review.openstack.org/336308 | 14:15 |
| psachin | rcernin, No not like swiftonfile. Say I have one nfs mount point(/mnt/nfs/m1), can I configure swift to use that is one of my device node ? | 14:16 |
| openstackgerrit | Maria Malyarova proposed openstack/swift: Added missing parenthesis in print calls https://review.openstack.org/337690 | 14:16 |
| tdasilva | psachin: I don't think you can use nfs because swift requires xattr support, and yes you are correct, swiftonfile is a bit different | 14:17 |
| tdasilva | swifonfile also requires a filesystem with xattr support, so NFS is also not supported there | 14:17 |
| *** hseipp has quit IRC | 14:18 | |
| psachin | Cools! Thanx tdasilva. | 14:20 |
| *** dmellado has quit IRC | 14:21 | |
| *** dmellado has joined #openstack-swift | 14:21 | |
| rcernin | What would be the minimum hardware requirements to run swift node as standalone ? | 14:23 |
| psachin | rcernin, any commodity h/w | 14:23 |
| *** sheel has joined #openstack-swift | 14:24 | |
| klamath | rcernin: the last "best practice" on sizing for swift is 1 CPU thread per 2 physical drives | 14:24 |
| *** ManojK has joined #openstack-swift | 14:25 | |
| klamath | proxy tier is super cpu/network heavy | 14:25 |
| *** oshritf has joined #openstack-swift | 14:25 | |
| tdasilva | rcernin: HP gave a talk on their public cloud a while back, it has some data there on what kind of hardware they were using: https://www.openstack.org/summit/vancouver-2015/summit-videos/presentation/maintaining-and-operating-swift-at-public-cloud-scale | 14:27 |
| *** dmellado has quit IRC | 14:29 | |
| *** dmellado has joined #openstack-swift | 14:30 | |
| *** ManojK has quit IRC | 14:53 | |
| *** ManojK has joined #openstack-swift | 14:55 | |
| *** diogogmt has joined #openstack-swift | 14:58 | |
| *** psachin has quit IRC | 14:59 | |
| *** nadeem has joined #openstack-swift | 15:04 | |
| *** Jeffrey4l_ has quit IRC | 15:11 | |
| *** oshritf has quit IRC | 15:17 | |
| *** openstackgerrit has quit IRC | 15:33 | |
| *** openstackgerrit has joined #openstack-swift | 15:34 | |
| *** gyee has joined #openstack-swift | 15:45 | |
| *** jmccarthy has quit IRC | 15:49 | |
| *** jmccarthy has joined #openstack-swift | 15:50 | |
| *** takashi has joined #openstack-swift | 16:09 | |
| *** ManojK has quit IRC | 16:09 | |
| *** adu has joined #openstack-swift | 16:14 | |
| *** baojg has joined #openstack-swift | 16:17 | |
| *** takashi has quit IRC | 16:17 | |
| *** vint_bra has joined #openstack-swift | 16:26 | |
| *** klrmn has joined #openstack-swift | 16:26 | |
| *** rcernin has quit IRC | 16:31 | |
| *** asettle has quit IRC | 16:32 | |
| *** tesseract- has quit IRC | 16:33 | |
| *** pcaruana has quit IRC | 16:33 | |
| *** hogepodge has quit IRC | 16:34 | |
| *** CR3ZNEZCH has joined #openstack-swift | 16:39 | |
| *** rledisez has quit IRC | 16:39 | |
| openstackgerrit | Maria Malyarova proposed openstack/swift: Added missing parenthesis in print calls https://review.openstack.org/337690 | 16:40 |
| *** sbelous_ has joined #openstack-swift | 16:40 | |
| *** hogepodge has joined #openstack-swift | 16:41 | |
| timburke | good morning | 16:43 |
| *** ManojK has joined #openstack-swift | 16:44 | |
| *** baojg has quit IRC | 16:46 | |
| *** sbelous_ has quit IRC | 16:48 | |
| *** hogepodge has quit IRC | 16:50 | |
| *** Suyash has joined #openstack-swift | 16:56 | |
| *** mmcardle has quit IRC | 16:56 | |
| *** adu has quit IRC | 17:00 | |
| *** hogepodge has joined #openstack-swift | 17:02 | |
| *** permalac has joined #openstack-swift | 17:09 | |
| *** adu has joined #openstack-swift | 17:13 | |
| *** tqtran has joined #openstack-swift | 17:14 | |
| *** rcernin has joined #openstack-swift | 17:16 | |
| *** hogepodge has quit IRC | 17:17 | |
| *** hogepodge has joined #openstack-swift | 17:23 | |
| *** adu has quit IRC | 17:24 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Have py35 tox env match py34 https://review.openstack.org/337802 | 17:24 |
| *** hogepodge has quit IRC | 17:28 | |
| *** ManojK has quit IRC | 17:30 | |
| openstackgerrit | Alistair Coles proposed openstack/swift: Check tmp_dir does not exist before calling mkdirs https://review.openstack.org/337804 | 17:33 |
| *** hogepodge has joined #openstack-swift | 17:36 | |
| *** acoles is now known as acoles_ | 17:44 | |
| *** baojg has joined #openstack-swift | 17:47 | |
| *** johanfi has joined #openstack-swift | 17:48 | |
| johanfi | hi.. I've got some issues with connecting keystone and swift (Ubuntu 16.04 - Mitaka) | 17:48 |
| johanfi | if I try to use the swift-client , it seems to default to use v2 of the keystone API, not v3 (which is the one we are using) | 17:49 |
| johanfi | i tried both using the --os-variables and using the textfile with sourcing | 17:50 |
| johanfi | authorization failed, the request you have made requires authentication | 17:51 |
| *** asettle has joined #openstack-swift | 17:51 | |
| johanfi | by using tokens, not passwords I get Excpecting to find username or UserId | 17:53 |
| zaitcev | Just don't default, then. Pass -A 3. | 17:53 |
| *** baojg has quit IRC | 17:53 | |
| johanfi | keystone still says I'm using v2 | 17:54 |
| timburke | johanfi: i think zaitcev meant -V 3 | 17:55 |
| *** hogepodge has quit IRC | 17:55 | |
| johanfi | I tried both | 17:55 |
| johanfi | :) | 17:55 |
| johanfi | but no success | 17:56 |
| *** ManojK has joined #openstack-swift | 17:58 | |
| timburke | johanfi: if you include a --debug flag, it should include lines referring to either keystoneclient.auth.identity.v3.base or keystoneclient.auth.identity.v2 -- did it use v2 both with and without the -V 3? | 18:01 |
| *** CaioBrentano has joined #openstack-swift | 18:01 | |
| *** CR3ZNEZCH has quit IRC | 18:03 | |
| johanfi | how does the proxy-server know which version of keystone-api to use? | 18:04 |
| johanfi | does it expect the client to know? | 18:05 |
| timburke | proxy-side, it's in the config for the authtoken middleware. around https://github.com/openstack/swift/blob/master/etc/proxy-server.conf-sample#L301 | 18:06 |
| clayg_ | is patch 336741 really just being that unlucky? | 18:06 |
| patchbot | clayg_: https://review.openstack.org/#/c/336741/ - swift - Add basic functests for user object metadata | 18:06 |
| timburke | looks like we need to add instructions regarding API version, though :/ | 18:06 |
| clayg_ | oh i think it was just setuptools 24.0.0 - KeyError processor_architecture | 18:07 |
| *** clayg_ is now known as clayg | 18:07 | |
| timburke | johanfi: from http://docs.openstack.org/developer/keystonemiddleware/middlewarearchitecture.html#configuration it looks like you can specify auth_version in the proxy-server.conf | 18:07 |
| johanfi | string value = v3 | 18:09 |
| johanfi | ? | 18:09 |
| zaitcev | This may be a bit worrysome, but when they moved options from __init__.py to _auth.py of keystonemiddleware's auth_token, they auth_version disappeared. | 18:09 |
| zaitcev | See for yourself | 18:09 |
| zaitcev | https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_auth.py#L147 | 18:09 |
| zaitcev | Oh, it's in another _OPTS now - https://github.com/openstack/keystonemiddleware/blob/master/keystonemiddleware/auth_token/_opts.py#L50 | 18:11 |
| johanfi | I see | 18:11 |
| johanfi | but it is a string value? what is the correct name for v3 | 18:12 |
| timburke | judging by their tests, "v3" should work | 18:12 |
| *** timss has joined #openstack-swift | 18:13 | |
| *** adu has joined #openstack-swift | 18:13 | |
| timburke | i'd also expect that you should be able to leave the endpoint unversioned and let keystonemiddleware handle the negotiation, though.... | 18:13 |
| johanfi | tried both v3.0 and v3, still uses v2.0 when communicating with keystone | 18:15 |
| -openstackstatus- NOTICE: Job instability resulting from a block storage connectivity error on mirror.iad.rax.openstack.org has been corrected; jobs running in rax-iad should be more reliable again. | 18:15 | |
| zaitcev | should not matter, they are all passed to a normalizer that drops v https://github.com/openstack/keystoneauth/blob/master/keystoneauth1/discover.py#L73 | 18:15 |
| timburke | johanfi: (we've determined that it's definitely the proxy<->keystone communication that's problematic rather than client<->keystone, yes?) | 18:19 |
| *** adu has quit IRC | 18:20 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/swift: Updated from global requirements https://review.openstack.org/88736 | 18:20 |
| johanfi | I think so | 18:20 |
| timburke | johanfi: i'll see if i can reproduce the trouble. what version of keystonemiddleware are you using? | 18:23 |
| johanfi | 4.4.0-3 | 18:25 |
| timburke | in the mean time, if you want to verify that the client can talk to keystone just fine, you could try running `swift auth` -- it should auth with keystone then print a storage URL and token, without actually talking to swift at all | 18:25 |
| *** hogepodge has joined #openstack-swift | 18:26 | |
| johanfi | that works | 18:26 |
| johanfi | but I had to give IDs, not usernames | 18:27 |
| openstackgerrit | Merged openstack/python-swiftclient: Add __ne__ built-in function https://review.openstack.org/337427 | 18:28 |
| clayg | cschwede: briancline: I'd like to rebase patch 335373 - also review called from some additional tests - can I do those too? | 18:30 |
| patchbot | clayg: https://review.openstack.org/#/c/335373/ - swift - Prevent down nodes failing PUTs on utf8 obj names | 18:30 |
| clayg | i bumped lp bug #1597210 to critical, we should fix it before a new release - 500's sux | 18:31 |
| openstack | Launchpad bug 1597210 in OpenStack Object Storage (swift) "Downed object-server during a PUT with unicode characters causes UnicodeDecodeError and 500 status code" [Critical,In progress] https://launchpad.net/bugs/1597210 - Assigned to Brian Cline (briancline) | 18:31 |
| *** asettle has quit IRC | 18:37 | |
| openstackgerrit | Merged openstack/swift: pickle_async_update should create tmp_dir https://review.openstack.org/330363 | 18:40 |
| clayg | oh the conflict is sorta nasty with the refctoring - easier to just re-apply the fix I think | 18:41 |
| torgomatic | clarkb: true that there's an XFS filesystem in the functional tests, but gate-swift-tox-func runs on ext4, and it seems to be running out of xattr space on some tests on a patchset I proposed | 18:43 |
| clarkb | torgomatic: is tox-func not the functional tests with an XFS fs? | 18:43 |
| torgomatic | clarkb: there's tox-func which runs them in an in-process mode (on ext4), and then swift-dsvm-functional which runs them against a real proxy over a real socket (presumably on XFS since it works :) ) | 18:44 |
| clarkb | gotcha | 18:45 |
| torgomatic | there was a mock of getxattr / setxattr that I'm trying to get rid of since it has some bad behavior | 18:45 |
| torgomatic | (notably, if you unlink a file with real xattrs and then create a new file with the same name, the xattrs are gone... not so with this mock) | 18:46 |
| torgomatic | well, *really* I'm trying to add xattr checksumming since filesystems don't provide this | 18:46 |
| timburke | johanfi: did you use ids or names in proxy-server.conf? i'm seeing something similar when trying to use discovery; it seems to try v2 auth (for the service user) if i just specify project name/username/password, but will do v3 if i additionally specify project_domain_name and user_domain_name | 18:46 |
| clayg | i'm confused about that bug - is _ supposed to return a unicode object or a bytestr? | 18:48 |
| timburke | johanfi: even then, though, it will use the v2 endpoint to try to validate the token received from the client, however. looks like this is due to how the service user's catalog is configured though (the keystone server i'm using has it returning v2 identity endpoints even in a v3 response), so it's probably worth checking that | 18:48 |
| *** baojg has joined #openstack-swift | 18:50 | |
| timburke | clayg: or native string? i.e., bytes in py2 and unicode in py3 -- i have no idea | 18:50 |
| *** hogepodge has quit IRC | 18:53 | |
| *** baojg has quit IRC | 18:54 | |
| *** hogepodge has joined #openstack-swift | 19:00 | |
| openstackgerrit | Maria Malyarova proposed openstack/swift: Added missing parenthesis in print calls https://review.openstack.org/337690 | 19:22 |
| *** sheel has quit IRC | 19:25 | |
| *** hogepodge has quit IRC | 19:27 | |
| *** hogepodge has joined #openstack-swift | 19:29 | |
| *** gyee has quit IRC | 19:31 | |
| *** nadeem has quit IRC | 19:37 | |
| clayg | jrichli: that looks bad! the response inclued X-Object-Transient-Sysmeta!? | 19:38 |
| jrichli | clayg: I debugged that sort of late at night, and IIRC, the 2 dicts were equal when examining in debugger - even when reproducing the issue. So, I am gonna have to try to reproduce again and see whats up | 19:40 |
| jrichli | clayg: but yes, I saw that in the assert failure | 19:41 |
| clayg | from https://pythonhosted.org/kitchen/unicode-frustrations.html#frustration-6-inconsistent-apis-part-deux -> Note kitchen.i18n provides alternate gettext translation objects that return only byte str or only unicode string. | 19:41 |
| clayg | because apparently gettext is a mixed bag - hooray! | 19:42 |
| clayg | jrichli: ok, np - you migth give it a chance to pass in the gate - it seemed to be a provisiong error - i'm not sure if -infra is backed up or what? | 19:42 |
| *** hogepodge has quit IRC | 19:43 | |
| *** hogepodge has joined #openstack-swift | 19:44 | |
| *** baojg has joined #openstack-swift | 19:50 | |
| *** joeljwright has joined #openstack-swift | 19:51 | |
| *** ChanServ sets mode: +v joeljwright | 19:51 | |
| *** baojg has quit IRC | 19:55 | |
| *** ManojK has quit IRC | 19:57 | |
| *** asettle has joined #openstack-swift | 19:58 | |
| *** asettle has quit IRC | 19:58 | |
| *** ManojK has joined #openstack-swift | 19:58 | |
| *** sheel has joined #openstack-swift | 20:01 | |
| *** CaioBrentano has quit IRC | 20:07 | |
| *** vint_bra has quit IRC | 20:16 | |
| *** hogepodge has quit IRC | 20:29 | |
| clayg | i don't get it - everything is telling me that the strings returned from gettext are bytestr | 20:30 |
| clayg | swob seems to think it likes path to be utf-8 - which makes sense, and seems consitent with briancline's patch to "decode" the utf-8 str to unicode | 20:30 |
| clayg | but I don't understand why interpolating a unicode str into a bytestr wouldn't also blow up - or maybe more importantly why interpolating a bytestr into a bytestr blew up in the first place | 20:31 |
| *** hogepodge has joined #openstack-swift | 20:31 | |
| clayg | unless somehow on briancline's machines gettext returns unicode? but *why* | 20:32 |
| timburke | clayg: when did we see bytestr-into-bytestr blow up? | 20:32 |
| clayg | well you can't see it | 20:32 |
| clayg | his fix seems to indicate that his lhs was a unicode string - but mine aren't - no matter what locale I'm testing with - and he said he was testing with en.utf-8 | 20:33 |
| clayg | so... I have a theory | 20:41 |
| clayg | briancline: where you at!? | 20:41 |
| clayg | i would have *swore* this one time we had this crappy but with a unicode device name from the ring making things suck - but I can't think of what I might grep around for | 20:43 |
| clayg | this exception_occurred path is sort of interesting in that we gettext a template with "%(info)s" then interpolate another gettext'd message into that template | 20:45 |
| *** hogepodge has quit IRC | 20:45 | |
| clayg | looking at the traceback more carefully it didn't blow up interpolating the template or "additional_inf" - it blew up putting them together | 20:45 |
| openstackgerrit | Romain LE DISEZ proposed openstack/swift: Permit to bind object-server on replication_port https://review.openstack.org/337861 | 20:46 |
| clayg | the additional_info included utf-8 chars, but the template was unicode - probably not because the original gettext'd template *returned* unicode, but because 'my str %s' % arg returns unicode if arg is a unicode str (even if it doesn't have non-ascii chars in it) | 20:46 |
| *** hogepodge has joined #openstack-swift | 20:47 | |
| timburke | clayg: yes! definitely true | 20:47 |
| clayg | if I could remmeber how to get my devices in my ring to be unicode I think I could even prove it | 20:47 |
| timburke | i saw something similar in the rsync stuff cschwede was working on | 20:47 |
| timburke | i think it involves a json serialization/deserialization cycle? | 20:48 |
| clayg | oh maybe, iirc that was just the gettext'd string returning non-ascii chars in utf-8 and then you can't interpolate the unicode str into it - if the string came from json it'd be unicode | 20:49 |
| clayg | oh... the ring bits *are* json - so always unicode | 20:51 |
| clayg | yeah this code path is fucked | 20:51 |
| *** baojg has joined #openstack-swift | 20:51 | |
| clayg | jesus hell | 20:51 |
| timburke | yeah. the unicode / bytes dichotomy sucks | 20:52 |
| *** vint_bra has joined #openstack-swift | 20:54 | |
| *** gyee has joined #openstack-swift | 20:55 | |
| *** baojg has quit IRC | 20:55 | |
| clayg | does NOT blow up -> 'test %(arg1)s %(arg2)s' % {'arg1': 'wat', 'arg2': 'w\xc8\x81t'} | 20:56 |
| clayg | DOES blow up -> 'test %(arg1)s %(arg2)s' % {'arg1': u'wat', 'arg2': 'w\xc8\x81t'} | 20:56 |
| clayg | becuase *obviously* the return type there should be unicode | 20:57 |
| timburke | clayg: also does NOT blow up: 'test %(arg1)s %(arg2)s' % {'arg1': 'wat', 'arg2': u'w\u0201t'} | 20:58 |
| clayg | holy crap, how long as this been broken for downed nodes when uploading objects with non-ascii chars | 20:59 |
| timburke | probably since about forever | 20:59 |
| timburke | that just seems to be the general trend with lurking unicode bugs :-/ | 20:59 |
| clayg | no, I think it would have been like sometime around fixing json? i guess that's basically forever :'( | 20:59 |
| timburke | oh, yeah. i suppose it's technically my fault :-( | 21:00 |
| timburke | or maybe i'm just thinking of swiftclient? https://review.openstack.org/#/c/184147/ | 21:01 |
| patchbot | timburke: patch 184147 - python-swiftclient - Remove simplejson dependency (MERGED) | 21:01 |
| *** sgundur1 has joined #openstack-swift | 21:03 | |
| timburke | looks like it was a series of four commits that did the conversion, stretched out between 2.4.0 and 2.6.0 | 21:03 |
| *** sgundur1 has left #openstack-swift | 21:03 | |
| *** asettle has joined #openstack-swift | 21:07 | |
| *** hogepodge has quit IRC | 21:09 | |
| *** hogepodge has joined #openstack-swift | 21:14 | |
| *** nadeem has joined #openstack-swift | 21:18 | |
| *** nadeem has quit IRC | 21:21 | |
| *** nadeem has joined #openstack-swift | 21:25 | |
| *** NM has quit IRC | 21:31 | |
| *** NM has joined #openstack-swift | 21:31 | |
| *** NM has quit IRC | 21:36 | |
| *** baojg has joined #openstack-swift | 21:52 | |
| *** _JZ_ has quit IRC | 21:54 | |
| *** baojg has quit IRC | 21:56 | |
| *** ManojK has quit IRC | 21:56 | |
| *** ManojK has joined #openstack-swift | 21:59 | |
| openstackgerrit | Tim Burke proposed openstack/swift: Fix gettext_ calls https://review.openstack.org/336743 | 21:59 |
| *** ManojK has quit IRC | 22:06 | |
| *** vint_bra has quit IRC | 22:10 | |
| *** adu has joined #openstack-swift | 22:18 | |
| *** asettle has quit IRC | 22:24 | |
| *** sheel has quit IRC | 22:45 | |
| *** baojg has joined #openstack-swift | 22:52 | |
| mattoliverau | morning | 22:52 |
| *** nadeem has quit IRC | 22:57 | |
| *** baojg has quit IRC | 22:57 | |
| *** ManojK has joined #openstack-swift | 22:59 | |
| *** nadeem has joined #openstack-swift | 23:03 | |
| *** adu has quit IRC | 23:05 | |
| zaitcev | Great, I have to hack on the kernel a little bit now. Thanks, Sam. | 23:08 |
| *** kei_yama has joined #openstack-swift | 23:13 | |
| *** rcernin has quit IRC | 23:18 | |
| *** ManojK has quit IRC | 23:23 | |
| *** joeljwright has quit IRC | 23:27 | |
| *** sambhare has joined #openstack-swift | 23:33 | |
| *** sambhare has quit IRC | 23:33 | |
| openstackgerrit | Clay Gerrard proposed openstack/swift: Prevent down nodes failing PUTs with non-ascii obj names https://review.openstack.org/335373 | 23:34 |
| *** permalac has quit IRC | 23:35 | |
| *** ManojK has joined #openstack-swift | 23:41 | |
| clayg | i'm starting to think it'd be easier to just let gettext strings be utf-8, let req.path be utf-8, and just make the damn device/node args from the ring *also* be utf-8 since they probably almost always include only ascii chars :'( | 23:52 |
| clayg | it's amazing we dont' have more of these bugs than we do :'( | 23:53 |
| *** baojg has joined #openstack-swift | 23:53 | |
| *** hosanai has joined #openstack-swift | 23:55 | |
| *** ChanServ sets mode: +v hosanai | 23:55 | |
| briancline | clayg: yea i'm here, busy day | 23:55 |
| clayg | briancline: s'ok - i figured it out w/o you | 23:56 |
| clayg | well - i know what the *problem* is - not sure about the solution - more than one way to fix it | 23:56 |
| *** baojg has quit IRC | 23:58 | |
| briancline | i found another exception elsewhere btw, planning to push up another revision with that and the extra tests | 23:58 |
| briancline | oh? what'd you find? | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!